Obscuration of quasars by dust in damped Lyman-alpha systems

We present a largely analytical method to compute the obscuration of quasars by dust in damped Lyα absorption systems. The inputs required by our method are the observed luminosity function of quasars and the typical dust-to-gas ratio and empirical distribution of H I column densities in the damped Lyα systems. We use the dust-to-gas ratio inferred from the reddening of background quasars in our previous studies. As outputs, we obtain self-consistent corrections for the effects of obscuration on the luminosity function of quasars and the distribution of H I column densities in the damped Lyα systems. Our calculations indicate that 10%-70% of the bright quasars at z = 3 are missing from optical samples. This is not enough obscuration to explain the flattening or turnover in the observed comoving density of bright quasars in the interval 2 <~ z <~3. It does, however, help to account for the ultraviolet background radiation implied by the proximity effect at z ~ 3 We estimate that quasars produce at least 10% and possibly all of this radiation. According to our extrapolations the obscuration increases so rapidly with redshift that samples of optically selected quasars may be only 10% complete at z = 4. The uncertainties in our estimates of the obscuration are caused mainly by the weak constraints on the number of damped Lyα systems with the highest H I column densities. For the same reason, we can only set weak constraints on the comoving densities of H I and dust in the damped Lyα systems: 1 X 10^-3^ <~hOMEGAH I_ <~ 3 X 10^-2^ and 10^-6^ <~hOMEGAdust_ <~ 10^-4^. The limits from COBE on the far-infrared background radiation imply that the dust in the damped Lyα systems cannot be much hotter than the dust in the Milky Way.